This case has only been checked with OpenFOAM v9.

To run the case first execute "setupCase" this script will copy a timestep from a sorted folder together with the mesh. A snappyHexMesh routine is commented in the "setupCase" script, this can be used if modifications to the mesh is desired.

First the case has to be run until the flow has reached the outlet with Cc and Cv set to 0 in transportProperties. A timestep for this is provided in the storage folder, but also copied into the case by the "setupCase" script.
After this Cc and Cv has to brought up in value slowly, as a too big increase will crash the simulation. Increments of 0.05 were used.
When Cc and Cv have been increased in value, the averaging must be started. There is a function for the field averaging in controlDict, that can be commented in for this purpose. The averaging time of the results presented in the paper were 0.002 s

Please check the write settings in controlDict, as they are most likely not optimal for an environment outside of where the case that was originally run.

Images in the paper have been made either with a iso volume of alpha.water 0.49-0.51 or by taking a plot over line of the average velocity field in the centerline of the nozzle throat at 1.5, 3 and 6 mm into the throat from the inlet side.

The simulation might be more stable if there are no layers on the "walls_exit" patch. A mesh without this can be generated by removing the "walls_exit" entry on line 257 of snappyHexMeshDict.layers, cleaning the case and running "setupCase" with the snappyHexMesh routine. This is only a theory and has not been tested by the author.
